In the last several decades, analyses of the neurobiological sequelae of the aging process has led to increasingly sophisticated understanding of both the extent and the specificity of age-related changes in brain structure and physiology. In spite of these achievements, there remains a relatively poor understanding of which age-related biological changes directly underlie the accompanying loss of functional capacity, and why. During the same era, there has been a revolution in the conceptual foundations of cognitive science. This revolution has lead to the development of network models at various levels of abstraction which have achieved extraordinary success in accounting for the fundamental data of human and animal cognition, in particular, associative memory processes. Surprisingly, there has been little attempt to apply this powerful conceptual framework to the design of experiments that might shed light on the manner in which neurobiological alterations contribute to functional impairments in the aging brain. It is the intent of the proposed research to begin to bridge the gap between these two disciplines with experiments that exploit recent advances in neurophysiological recording methods, permitting simultaneous recording from large populations of neuronal elements, within a conceptual framework provided by the theory of attractor neural networks and autoassociative pattern completion. Specifically, we will use the method of cross-correlation analysis to investigate the magnitude and duration of experience-dependent changes in synaptic coupling in young and old animals, and to test the hypotheses that synaptic loss during aging creates conditions of instability of neuronal cell assemblies. In addition, we will use temporary inactivation of hippocampal commissural connections to test the theory that the age-related loss of neurobiological resources places it near the critical point of the graceful degradation function that characterizes the initially slow, but subsequently catastrophic manner in which increasing loss of synaptic connections leads to functional degradation in neural networks. Finally, we will investigate the possibility that neural activity related to an impending goal is maintained in the hippocampus during a delayed conditional spatial response task, and how the quality of such goal representations change with age.